From the recent discussion above it would appear that applying the official MS fix on the server is likely to create considerable IT waves with mapped drives being lost and presumably having to be remapped. I'm assuming that the remapping would have to be
done on all client machines connected to the server.......not a small job if there are many users.
In light of that, it seems that the fix that has been employed successfully on the client machines only may be the less painful path which wouldn't introduce the problems with mapped drives or involve the IT department.
The downside I can see of the client side fix is that one new or re-imaged client machine that hasn't had the fix employed would be enough to break the database.
I'm trying to look at all angles regarding a database I'll be installing shortly for a new client and what method will cause the least disruption.